MBZUAI Highlights Innovative AI Projects At IROS And GITEX Global Events

Since its inception in 2019, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) has focused on advancing AI research and innovation. The university aims to tackle real-world issues in healthcare, sustainability, education, and smart cities. This week, two major technology events in the UAE showcase the progress made by MBZUAI's faculty, students, and researchers over the past five years.

Attendees at the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024) in Abu Dhabi and the annual GITEX Global technology event in Dubai are exploring new projects from MBZUAI. These initiatives highlight the university's commitment to AI-driven solutions across various fields.

One of the showcased projects is FRAPPE (FRAming, Persuasion, and Propaganda Explorer). This system analyses media content for framing, persuasion, and propaganda techniques. It offers real-time analysis of articles based on genre, framing, and persuasion strategies. This allows comparisons across different media outlets, countries, and languages.

FRAPPE aims to enhance media literacy by exposing techniques used to shape opinions and emotions. By promoting informed decision-making and understanding of media manipulation, it helps build trust in media. The tool supports transparent information dissemination worldwide.

Another innovative project is BiMediX2 – an AI Virtual Arabic Doctor. It is a bilingual medical large language model (LLM) and large multimodal model (LMM). BiMediX2 can understand and generate content across text, images, audio, and video data types.

This model facilitates various medical interactions like multi-turn chats and question answering. It can also analyse medical images such as x-rays, CT scans, and MRIs. The research team sees its potential use in remote areas where telemedicine platforms connect doctors with patients.

The team is expanding BiMediX2's functionality to locate nearby physicians and facilities. It will also book appointments for patients needing further medical engagement. This development could significantly improve healthcare access in underserved regions.

Dog Bot: Enhancing Public Services

The 'Dog Bot' project involves a visual analytics generative pre-trained transformer (GPT). It combines image, text, and audio capabilities for speech recognition and understanding. This solution could be applied across sectors like health inspections at work sites or assisting police patrols.

Dog Bot can survey agricultural sites for crop performance or interact with patients remotely by sharing diagnostic images with doctors. Essentially, it acts as the eyes and ears of crucial public services.

Sustainability Efforts with GeoChat+

MBZUAI researchers are addressing sustainability issues through GeoChat+. This vision-language model (VLM) focuses on multi-modal temporal remote sensing image analysis to monitor deforestation effectively.

GeoChat+ integrates data from Earth observation systems like Sentinel-1 and Sentinel-2 along with Landsat imagery from WorldView or Google Earth sources. Its goal is providing comprehensive geospatial insights for applications ranging from environmental monitoring to disaster response planning or urban development strategies.

The university's efforts demonstrate a strong commitment towards using AI technologies for solving pressing global challenges while fostering innovation within diverse domains such as healthcare or sustainability initiatives alike.

With inputs from WAM